Article
WPFD 2022: Addressing digital challenges to freedom of expression through UNESCO’s Recommendation on Ethics of AI
The plenary was moderated by Natalia Zuazo, Consultant of the Communication and Information Sector for UNESCO Regional Office for Latin America and the Caribbean. Zuazo established the complementarities between the Windhoek+30 Declaration and the Recommendation on the Ethics of AI, two instruments that facilitate the interaction between the communities of practice. “These tools share common grounds in promoting transparency, accountability, and more open processes for multi-stakeholder decision-making, while developing capacity building at different levels”, she said.
Gabriela Ramos, Assistant Director-General for Social and Human Sciences at UNESCO set the stage by presenting the Recommendation on the Ethics of AI and how the values, principles and the policy areas of the Recommendation address privacy and freedom of expression issues, and proposed the need for a new business model for data management and protection.
Given that AI has far-reaching impact on all fundamental freedoms, including the freedom of expression, I want to underscore that self-regulation is insufficient to control the downsides of AI. Such voluntary standards could easily be disregarded in the pursuit of profits. We believe that policies matter, regulatory frameworks matter, and they can achieve positive outcomes by balancing strong protection with allowing the freedom of expression to flourish.
Challenges and opportunities of AI and freedom of expression
Tools like AI can be used for good and to enhance freedom of expression and to enhance freedom of press, said Fabrizio Scrollini, Executive Director of Latin American Initiative for Open Data (ILDA). “We are seeing newsrooms using AI to enhance their analytic abilities, exploring corruption cases, and making sense of large sets of documents that wouldn't be possible to actually understand without the help of AI,” he said.
However, he acknowledged that there are challenges like data governance that prevent freedom of expression from thriving with AI use. “These beautiful algorithms are as good as the data they get,” he said, “But we don’t get good data because our biases and our history prevent us.”
Similarly, Alfonso Peralta Gutierrez, Judge of the First Instance and Criminal Investigation Court n. 1 of Roquetas de Mar, Spain, said that “We must know that AI is like a knife. It has its pros and cons.”
It can boost freedom of expression, but at the same time, it can damage the freedom of expression. It can be used to create fake news, disinformation, or deep fakes, but also it can help to identify them.
Rumman Chowdhury, Director for Machine Learning Ethics, Transparency and Accountability at Twitter, noted that content curation is happening on any sort of digitally mediated platform. Chowdhury underlined that to responsibly use AI, one needs to be aware where content is created and which tech narratives are overemphasized. Moreover, there is a need to emphasize voices to ensure that more people can be heard.
While calling for a global debate on the role of digital platforms, Rodica Ciochina, Programme Officer in the Media and Internet Governance Unit of the Information Society Department of the Directorate General for Human Rights and Rule of Law, Council of Europe said, “Digital platforms need to realize the power that they carry, and that the responsibility comes with that as well as in shaping the public discourse and the consequences. We must not forget that in today's complex digital ecosystem, it is a shared responsibility, both for public and private actors and users to contribute to making full use of the benefits of innovation.”
Key takeaways from the session
Apart from a full implementation of the UNESCO Recommendation on the Ethics of AI, the panel also discussed other ways to address the challenges of AI on the freedom of expression.
Firstly, there is a need to organize communities and give them a voice, using smart design to help data regimes reflect inclusive and democratic values. According to Scrollini, “the fact is that we are not going to be heard unless we organize, and make sure we have every party in the room.”
Secondly, from a judicial perspective, judges and legal practitioners around the world must have a basic understanding of AI applications in the judicial context. For example, UNESCO’s Massive Open Online Course (MOOC) on AI and the rule of law trains judicial operators worldwide to understand AI initiatives in judicial systems around the world.
Finally, online platforms must balance protecting people’s rights of privacy and security with the need for transparency and accountability. Twitter, for example, has invested in initiatives such as an “algorithmic bias bounty” for open-source bias detection and privacy-enhancing technology.
